This is the only question I would like answered:
If A is a set of religious statements, then what can you tell me about any x, where x is an element of A?
I cannot think of anything which could be reliably inferred from the simple fact that A is a set of religious statements, or at least anything which might be useful in demarcating religious statements from nonreligious statements.
The same does not hold for science:
If B is a set of scientific statements, then what can you tell me about any x, where x is an element of B?
From this we can reliably infer that x has empirical content, or is nontautological i.e. falsifiable. That is the way we demarcate between scientific and nonscientific theories, in that scientific theories can be criticised in a manner which nonscientific theories cannot, that is by experiment and observation.
As far as I can tell, to say that set A is a set of religious statements tells us absolutely nothing about any element of A, since whatever criterion we adopt, such as 'not reality-based,' 'concerning the supernatural,' or 'nonscientific,' we end up permitting into set A, too much, or too little.
For example, 'nonscientific' permits far too much, since all of mathematics and metaphysics, including metascience would be classed as religoius. Whereas 'not reality-based' permits too little, since almost all traditional religions make very strong claims about what reality is like, indeed one of the hallmarks of such traditions is that all facts confirm the theory.
I am beginning to believe that the word "religion" is quite superfluous, and without any meaningful content. It has no significance except that which people treat it, but is otherwise quite arbitrary. Somewhat like the emporers new clothes, it carries around a mystique, as though it refers to something important, but upon closer scrutiny there simply does not seem to be anything there.
If this is true, then Gould's notion that religion and science occupy nonoverlapping majisteria is actually accurate, but only because religion has no majisteria to speak of. The guys at Reasons to Believe actually have it right in this regard, even if they have it severely and most spectacularly wrong in almsot every other.